UDoGeC: Essential Protein Prediction Using Domain And Gene Expression Profiles

With the advent of high - throughput technologies large amount of protein - protein interaction data are available. Many researchers studied this data and predicted the importance of essential proteins in disease diagnosis, cosmetic development and drug design. Knockout experiments consumed more time and money while predicting the essential protein and this motivated computational biologists to develop algorithms and mathematical model to predict essential proteins. Early algorithms were based on topological properties. However the major setback to these algorithms were the unreliability of protein data. To overcome this, newly developed algorithms tried to incorporate biological properties along with the topological properties. In this study we introduce a new algorithm called, UDoGeC, Unified Domain and Gene Expression Centrality Method, which combines both the domain and gene expression profiles together. This algorithm is based on the assumption that an essential protein tends to form densely populated clusters and these clusters are strongly co-expressed. If that protein has rarely occurring domains than in other protein we predict it as essential otherwise non - essential. Finally a comparison study with three other centrality methods DC, UDoNC and PeC is performed to evaluate the performance of this newly suggested algorithm. The results were promising that UDoGeC showed better performance in various aspects. (C) 2016 Published by Elsevier B.V.